198 THE SEINE, THE MEUSE, AND THE MOSELLE 
In this valley it receives two branches on the southern side, to 
which special attention should be given. The first is the Sur- 
melin, whose head is found in the upland near its eastern pre- 
cipitous margin ; but, curiously enough, although this stream of 
course diminishes toward its source near Montmort, the valley 
that it occupies maintains an almost constant width some six 
miles farther, nearly to the escarpment of the upland. The 
second branch is the Petit Morin. This, like the Marne, heads in 
the lowland east of the upland, and also, like the Marne, escapes 
by a deep and narrow valley through the upland. The lowland 
area that it drains is, however, very small, and for about ten 
miles from its head there is an extended marsh, known as the 
Marais de St. Gond, lying partly on the lowlands and partly in 
the entrance to the narrow valley in the U})land. 
In searching for a reason for this arrangement of the Marne 
and its two branches, it is important to notice that if the branches 
were prolonged eastward the}^ would both lead to streams, the 
Soude and the Somme,* flowing for some distance on the low- 
land toward the heads of the branches, but then turning north- 
ward and entering the Marne directl3^ 
The beheading of the Surmelin and the Petit — In explana- 
tion of all these facts let it be supposed that the two pairs, Soude- 
Surmelin and Somme-Morin, were once actuall}" continuous 
streams at a time before the lowland was eroded on the weak 
rocks east of the upland, and let the verity of the supposition be 
tested by the likelihood of a natural, spontaneous change from 
that condition to the present. 
When the paired streams flowed westward, they, like the Marne, 
must have run in the direction of the dip of the strata; hence 
they mav all be called consequent streams. They must all have 
passed from the weak Cretaceous strata to the resistant Tertiary 
strata. The INIarne is much the largest of these three streams, 
and its valley must have-been deepened rapidly, while the other 
valleys must have been deepened slowly. As the valleys were 
deepened they progressively widened, but the widening must 
have been much more rapid on the weak than on the resistant 
strata ; and the deep valle}'’ of the Marne must have widened in 
the weaker strata much more rapidl}" than the neighboring 
shallow valleys of the Soude-Surmelin and the Somme-Morin. 
Now the question arises, will the divides between these three 
valleys shift in such a manner as to alter the assumed original 
* Not to be confused with the river Somme in northwestern France. 
